AI Document Search (RAG)
Built with React, FastAPI, OpenAI, LangChain, FAISS
The Problem
Users needed to quickly find answers from large PDF documents without reading entire files. Traditional search is keyword-based and misses contextual meaning. RAG solves this by combining vector search with LLM-generated answers grounded in actual document content.
The Solution
An AI-powered document search system built with Retrieval-Augmented Generation (RAG). Users upload PDFs, which are chunked, embedded using OpenAI, and stored in a FAISS vector database. The LangChain pipeline retrieves relevant document chunks and generates accurate, context-grounded answers. Built with a React frontend and FastAPI backend for high-performance async processing.
Key Features
RAG-based document Q&A
PDF upload and processing
FAISS vector search
LangChain pipeline
Context-grounded responses
OpenAI embeddings
Technology Stack
Related Projects
Brain Blood Clot Detection
Deep learning model for detecting blood clots in brain CT scans using convolutional neural networks and image segmentation techniques.
TensorFlow, OpenCV, Python, Deep Learning
EpiScan — Cancer Detection
Cancer detection machine learning model achieving 92% accuracy using deep learning for medical image classification.
TensorFlow, Scikit-Learn, Python